1. Field of the Invention
The present invention relates to methods and apparatus for regulating lasers. More particularly, the present invention relates to methods and apparatus for regulating laser diodes.
2. Description of Related Art
Simulated emission from a GaAs semiconductor diode laser was first observed in the early 1960's. In the intervening years, the semiconductor diode laser (the "laser diode") has come to dominate the laser field in its technological importance. It has become the key element in an increasing number of applications, most notably in optical fiber communication and optical data storage. This "success" is due to the fact that semiconductor lasers can be simply pumped by passing a current through them at voltage and current levels that are compatible with those of integrated circuits, and because they can be modulated directly at frequencies in excess of 20 GHz. Laser diodes can also be conveniently made because they can be mass-produced by the same photolithographic techniques as electronic circuits and because they can be integrated monolithically with those circuits.
Laser diodes are not without shortcomings, however For example, laser diodes typically demonstrate strong non-linear characteristics for variations in operating temperature. Thus, the optical output intensity of a laser diode is difficult to regulate during variations in operating temperature. Laser diodes are also known to present considerable variations of the light power emitted over time, that is, as the laser diodes age.
To moderate the intensity of these variations, it is necessary to provide for regulation of the output light power. Processes are already known for regulating the light power emitted by a laser diode under continuous operating conditions in which an electric signal representative of the light power is compared with an electric reference signal corresponding to a reference power. These processes include means for generating an error signal if appropriate, and for automatically modifying the operating conditions of the laser diode upon generation of such an error signal.
Typically, prior art automatic power controls supply a bias current that is set at approximately the threshold current level of the laser diode. In this manner, when a pulse current is added to the bias current, the laser diode is effectively switched between a high and low optical output intensity corresponding to the wave form of the pulse current. The supplied bias current is controllably varied to be approximately at the threshold current level of the laser diode, particularly when variations in the threshold current level occur because of corresponding variations in the operating temperature of the laser diode.
For clarity and convenience in understanding the foregoing, a typical digital fiber optical communication system and the roles of laser diodes and specific regulators that may be found in such a system will now be discussed.
A typical digital fiber optical communication system has a transmitter at a first end and a receiver ar a second end. Between the transmitter and the receiver are fiber optical contacts, splices, and optical fiber The transmitter is an electro-optical interface which includes an amplifier, laser diode, and some kind of regulation or circuits for keeping the laser diode at the same working point. The receiver is an optoelectrical interface and consists of a photodiode, generally either of the positive-intrinsic-negative ("PIN") or avalanche photo diode ("APD") type, an amplifier and clock recovery circuitry.
It is important in digital optical fiber systems to have as good a power margin as possible. The power margin is the difference between upper and lower optical power limits. These upper and lower optical lower limits are typically set by the receiver. The upper limit is determined by distortion in the amplifier caused by overload. The lower limit, generally called sensitivity, is mainly determined by noise in the front end amplifier.
The upper limit gives the maximum output from the transmitter if there is no damping of the optical signal in the contacts, splices, and fiber. The lower limit tells how much power has to be left after the optical signal from the transmitter has been damped after passing a number of contacts and splices and kilometers of fibers. The upper and lower limits define a power budget. Recognizing that every component in an optical signal path causes some damping or penalty, it should be appreciated that only a limited number of contacts or splices or only a limited length of fiber can be included within any path before the power budget of that path is depleted. This power budget may also be decreased by tolerances in both the transmitter and receiver ends. The tolerance in the transmitter end is determined by quality of regulation of the optical output power. All of the aforementioned relations and penalties are measured in dBm where a penalty or damping first splice is defined as 10.times.log (Pin/Pout)dBm.
As previously mentioned, a laser diode is a nonlinear element and its characteristics are very dependent on working temperature and aging effects. For proper operation, it is necessary to work over the knee of the diode. If work takes place under the knee, the diode will have turn on delay and ringing effects which may be tolerable but only in small amounts. The ratio between maximum and minimum optical output power, the extinction ratio, should be as high as possible but is limited by maximum mean power and maximum turn on delay. Regulation has to adjust the working point as well as possible to compensate for temperature and aging delays, recognizing that any mismatch between the ideal and real values will cause a Penalty that will decrease the power budget.
One of the most common ways to regulate a laser is to have the laser diode work at a constant temperature by a regulated peltier element. Such regulators contain two regulation loops, one for the peltier element and one for the mean power regulation. The output power regulator senses the mean optical output power with the monitor pin diode, and adjusts the bias current to compensate for the by-age increased Ith or threshold current. In such cases changes in scope efficiency are quite small. This kind of regulation works well but is expensive to implement, has a high power consumption (that is, a couple of watts for temperature regulation), requires cooling, and consumes space. Further, the peltier element is not as reliable as the laser diode.
As previously mentioned there are a number of alternative methods to regulate a laser without temperature regulation. One such method is to have only mean power regulation. Such a method has a constant modulation current and a simple regulation of the bias current. This method gives a penalty of 5-6 dBm's for a PIN receiver and more for a APD receiver.
Another method, slightly more complex, is to complement a mean power regulator with a simple feed forward regulator for the modulation current. By sensing the temperature and predicting the modulation currents of the laser diode compensation for variations of slope efficiency can be made. This method is very simple but it requires matching between the sensing element and the characteristics of the laser and does not compensate for aging. This method gives a penalty of a couple of dBm's.
A third method is to regulate mean power as above and also to regulate the modulation power via some kind of low frequency modulated optical power. The optical output power is modulated with a low frequency signal, the amplitude of the signal being 10% or less of the data signal's amplitude so that it can be seen as an low frequency ("LF") ripple on the signal. Modulation can be done via the bias current and causes both maximum and minimum optical values to vary. If such modulation is done completely above the knee these variations would be the same for both maximum and minimum power, but if the bias current is decreased close to the knee or slightly under it, the minimum output power variations will be suppressed. This variation of the LF signal will be registered by the monitor and filtered away and used for regulation together with the mean power signal. This third method has a penalty of approximately a dBm caused by the LF Modulation. Additionally, regulation of this method can only work close to the knee for proper regulation. The optimal working point can at some times be out of this range.